Disturbance rejection is an important topic in control design since disturbances are inevitable in practical systems. To realise this target for nonlinear systems, this thesis brings in an assumption about the existence of a controlled invariant mani- fold and a Desired Feedforward Control (DFC) which is contained in the input to compensate the influence of disturbances. According to the approximation property of Neural Networks (NN) that any periodic signals defined in a compact set can be approximated by NN, the NN-based disturbance approximator is applied to approximate the DFC. Algorithmically, two important types of NN approximators that are Multi-layer Neural Networks (MNN) and Radial Basis Function Neural Networks (RBFNN) are presented in detail.In this thesis, a variety of nonlinear systems in standard canonical form are looked into. These forms are the output feedback form, the extended output feedback form, the decentralised output feedback form and the partial state feedback form. For these systems, four types of uncertainties are mainly considered. The first one is the disturbance that can be eliminated by the DFC. Secondly, the parameter uncertainty is taken into account. To get rid of this uncertainty, the adaptive control technique is employed for the estimation of unknown parameters, e.g. the NN gain matrix. The third one is the nonlinear uncertainty. For the case that nonlinear uncertainties are polynomials, it has a bound consisting of an unknown constant and a function of the regulated error such that this uncertainty can be also treated as the parameter uncertainty. Delay is the last type of uncertainty. Particularly, the delay is supposed to appear in output only. This uncertainty can be eliminated together with the nonlinear uncertainty. To establish the closed- loop stability, a Lyapunov-Krasovskii function is invoked. In addition, due to the requirement of the system structure or the stability analysis, some general control techniques are also involved such like the backstepping control and the high gain control.Throughout the results are illustrated by simulations.

No description available.

This study undertakes an empirical analysis of takeover predictions in the UK. The objectives of this research are twofold. First, whether it is possible to predict or identify takeover targets before they receive any takeover bid. Second, to test whether it is possible to improve prediction outcome by extending firm specific characteristics such as corporate governance variables as well as employing a different technique that has started becoming an established analytical tool by its extensive application in corporate finance field. In order to test the first objective, Logistic Regression (LR) and Artificial Neural Networks (ANNs) have been applied as modelling techniques for predicting target companies in the UK. Hence by applying ANNs in takeover predictions, their prediction ability in target classification is tested and results are compared to the LR results. For the second objective, in addition to the company financial variables, non-financial characteristics, corporate governance characteristics, of companies are employed. For the fist time, ANNs are applied to corporate governance variables in takeover prediction purposes. In the final section, two groups of variables are combined to test whether the previous outcomes of financial and non-financial variables could be improved. However the results suggest that predicting takeovers, by employing publicly available information that is already reflected in the share price of the companies, is not likely at least by employing current techniques of LR and ANNs. These results are consistent with the semi-strong form of the efficient market hypothesis.

No description available.

No description available.

No description available.

No description available.

No description available.

No description available.

Parallel simulation is a well developed technique for executing large and complex simulation models in order to obtain simulation output for analysis within an acceptable time frame. The main contribution of this thesis is the development of different adaptive techniques to improve the consistency, performance and resilience of the BSP Time Warp as a general purpose parallel simulation protocol. We first study the problem of risk hazards in the BSP Time Warp optimistic simulation protocols. Successive refinements to the BSP Time Warp protocol are carried out to eliminate errors in simulation execution due to different risk hazards. We show that these refinements can be incorporated into the BSP Time Warp protocol with minimal performance degradation. We next propose an adaptive scheme for the BSP Time Warp algorithm that automatically throttles the number of events to be executed per superstep. We show that the scheme, operating in a shared memory environment, can minimize computation load-imbalance and rollback overhead at the expense of incurring higher synchronization cost. The next contribution of this thesis is the study of different techniques for dynamic load-balancing and process migration for Time Warp on a cluster of workstations. We propose different dynamic load-balancing algorithms for BSP Time Warp that seek to balance both computation workload and communication workload, optimizing lookaheads between processors, as well as manage interruption from external workload. Finally, we propose an adaptive technique for BSP Time Warp that automatically varies the number of processors used for parallel computation based on the characteristics of the underlying parallel computing platform and the simulation workload.